Yolov4 Darknet weights to tflite - tensorflow

Question

I've successfully trained a Yolov4 Darknet detection model to detect lego bricks.
The Colab code is available here (with prefiled dataset): https://colab.research.google.com/drive/11d8BLJ4xnXrXumsV-0Z-jqPbz2cfWUUu
The generated config custom-yolov4-tiny-detector.cfg here: https://pastebin.com/68WLs2JM
And the file custom-yolov4-tiny-detector_finals.weights here: https://file.io/BeBokJG8G5lC
The test image produces awesome results with the following command:
!./darknet detect cfg/custom-yolov4-tiny-detector.cfg backup/custom-yolov4-tiny-detector_finals.weights {img_path} -dont-show
What I'd like to do now, is to export this model to TFlite.
I've tried to use the convertion code from https://github.com/hunglc007/tensorflow-yolov4-tflite.git
And especially that piece of code which create a temporary pb file:
!python save_model.py \
--weights /content/darknet/backup/custom-yolov4-tiny-detector_final.weights \
--output ./checkpoints/yolov4-tiny-pretflite-416 \
--input_size 416 \
--model yolov4 \
--tiny \
--framework tflite
The pretflite pb model is then converted to the final tflite:
%cd /content/tensorflow-yolov4-tflite
!python convert_tflite.py --weights ./checkpoints/yolov4-tiny-pretflite-416 --output ./checkpoints/yolov4-tiny-416.tflite
That function gives many error which make impossible to get a TFlite.
Traceback (most recent call last):
File "convert_tflite.py", line 76, in <module>
app.run(main)
File "/usr/local/lib/python3.8/dist-packages/absl/app.py", line 308, in run
_run_main(main, args)
File "/usr/local/lib/python3.8/dist-packages/absl/app.py", line 254, in _run_main
sys.exit(main(argv))
File "convert_tflite.py", line 71, in main
save_tflite()
File "convert_tflite.py", line 45, in save_tflite
tflite_model = converter.convert()
File "/usr/local/lib/python3.8/dist-packages/tensorflow/lite/python/lite.py", line 929, in wrapper
return self._convert_and_export_metrics(convert_func, *args, **kwargs)
File "/usr/local/lib/python3.8/dist-packages/tensorflow/lite/python/lite.py", line 908, in _convert_and_export_metrics
result = convert_func(self, *args, **kwargs)
File "/usr/local/lib/python3.8/dist-packages/tensorflow/lite/python/lite.py", line 1212, in convert
return self._convert_from_saved_model(graph_def)
File "/usr/local/lib/python3.8/dist-packages/tensorflow/lite/python/lite.py", line 1095, in _convert_from_saved_model
result = _convert_saved_model(**converter_kwargs)
File "/usr/local/lib/python3.8/dist-packages/tensorflow/lite/python/convert_phase.py", line 212, in wrapper
raise converter_error from None # Re-throws the exception.
File "/usr/local/lib/python3.8/dist-packages/tensorflow/lite/python/convert_phase.py", line 205, in wrapper
return func(*args, **kwargs)
File "/usr/local/lib/python3.8/dist-packages/tensorflow/lite/python/convert.py", line 809, in convert_saved_model
data = convert(
File "/usr/local/lib/python3.8/dist-packages/tensorflow/lite/python/convert.py", line 311, in convert
raise converter_error
tensorflow.lite.python.convert_phase.ConverterError: <unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_1/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_1/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_2/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_2/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_3/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_3/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_4/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_4/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_5/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_5/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_6/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_6/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_7/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_7/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_8/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_8/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_9/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_9/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_10/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_10/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_11/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_11/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_12/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_12/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_13/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_13/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_14/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_14/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_15/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_15/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_16/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_16/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_17/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_17/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_18/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): 'tf.FusedBatchNormV3' op is neither a custom op nor a flex op
<unknown>:0: note: loc(fused["StatefulPartitionedCall:", "StatefulPartitionedCall"]): called from
<unknown>:0: note: loc(callsite(callsite(fused["FusedBatchNormV3:", "model/batch_normalization_18/FusedBatchNormV3#__inference__wrapped_model_2529"] at fused["StatefulPartitionedCall:", "StatefulPartitionedCall#__inference_signature_wrapper_8064"]) at fused["StatefulPartitionedCall:", "StatefulPartitionedCall"])): Error code: ERROR_NEEDS_FLEX_OPS
<unknown>:0: error: failed while converting: 'main':
Some ops are not supported by the native TFLite runtime, you can enable TF kernels fallback using TF Select. See instructions: https://www.tensorflow.org/lite/guide/ops_select
TF Select ops: FusedBatchNormV3
Details:
tf.FusedBatchNormV3(tensor<?x104x104x32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<32xf32>) -> (tensor<?x104x104x32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
tf.FusedBatchNormV3(tensor<?x104x104x64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>) -> (tensor<?x104x104x64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
tf.FusedBatchNormV3(tensor<?x13x13x128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>) -> (tensor<?x13x13x128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
tf.FusedBatchNormV3(tensor<?x13x13x256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<256xf32>) -> (tensor<?x13x13x256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
tf.FusedBatchNormV3(tensor<?x13x13x512xf32>, tensor<512xf32>, tensor<512xf32>, tensor<512xf32>, tensor<512xf32>) -> (tensor<?x13x13x512xf32>, tensor<512xf32>, tensor<512xf32>, tensor<512xf32>, tensor<512xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
tf.FusedBatchNormV3(tensor<?x208x208x32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<32xf32>) -> (tensor<?x208x208x32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<32xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
tf.FusedBatchNormV3(tensor<?x26x26x128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>) -> (tensor<?x26x26x128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
tf.FusedBatchNormV3(tensor<?x26x26x256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<256xf32>) -> (tensor<?x26x26x256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<256xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
tf.FusedBatchNormV3(tensor<?x52x52x128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>) -> (tensor<?x52x52x128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<128xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
tf.FusedBatchNormV3(tensor<?x52x52x64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>) -> (tensor<?x52x52x64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<*xf32>) : {data_format = "NHWC", device = "", epsilon = 1.000000e-03 : f32, exponential_avg_factor = 1.000000e+00 : f32, is_training = false}
Is there a way to convert Yolov4 weights and config to a TFlite model ?